Re: [Ayatana-dev] libindicator API break

[Ted Gould <ted@xxxxxxxxxxxxx>]

On Fri, 2011-07-08 at 03:41 +0200, Marco Trevisan (Treviño) wrote:
> 2011/7/6 Ted Gould <ted@xxxxxxxxxxxxx>:
> > Hello,
> >
> > With the help of Trevino we've now dropped the multiple scroll signals
> > in libindicator.  This is a good thing, but it's forced an API break for
> > libindicator.  So those who are depending on it you'll need to use the
> > 0.4 version of the API here shortly.  No major change, just make sure
> > you're using the right scroll function, I think that Trevino has
> > provided patches for most of the users.

> I was just looking again at the code, but I noticed that also in my
> very first implementation (following what has already done on the
> default API), I didn't introduce any user_data pointer for the
> scrolled_entry signal. That is maybe not needed by userland (using
> libappindicator) since it was included there, but maybe it could be
> something we could introduce (also if it's always possible to attach
> some data to the Indicator object as a workaround)...
> In the indicator object interface that's the only signal not having an
> user_data field; I guess that it's something that some bindings
> (vala?!) would need.

It's actually added automatically by gsignal.  In reality the prototypes
aren't that big of a deal as gsignal uses it's own based on the
marshaller anyway.  The fun of function pointers in C!

		--Ted